You are on page 1of 16


Vol. 57, No. 10, October 2011, pp. 1721–1736
issn0025-1909 eissn1526-5501 11 5710 1721
Tournaments Without Prizes:
Evidence from Personnel Records
Jordi Blanes i Vidal, Mareike Nossol
Managerial Economics and Strategy Group, London School of Economics, London WC2A 2AE,
United Kingdom {,}
e use a quasi-experimental research design to study the effect of giving workers feedback on their relative
performance. The setting is a firm in which workers are paid piece rates and where, for exogenous reasons,
management begins to reveal to workers their relative position in the distribution of pay and productivity. We
find that merely providing this information leads to a large and long-lasting increase in productivity that is
costless to the firm. Our findings are consistent with the interpretation that workers’ incipient concerns about
their relative standing are activated by information about how they are performing relative to others.
Key words: tournaments; relative concerns; status concerns; relative performance feedback; relative
performance evaluation
History: Received November 4, 2009; accepted April 1, 2011, by Olav Sorenson, organizations. Published
online in Articles in Advance August 4, 2011.
1. Introduction
Does receiving feedback on relative performance
make workers strive for success, or does it lead
instead to dissatisfaction and loss of motivation? Does
this response depend on whether the comparison
with other workers is favorable or unfavorable? In
this paper we use a unique data set from a firm-level
quasi experiment to investigate these issues.
An important body of work has studied the use of
relative performance from the perspective of explicit
incentive theory (Lazear and Rosen 1981). Accord-
ing to tournament theory, an agent’s performance
can provide a useful benchmark upon which other
agents’ performance can be evaluated and rewarded.
The notion that competing with others for mone-
tary rewards can have motivation effects has received
empirical support in various economic settings (e.g.,
Knoeber and Thurman 1994, Eriksson 1999, Casas-
Arce and Martínez-Jerez 2009).
Alternatively, it may be that receiving feedback on
relative performance has in itself substantial conse-
quences for employee satisfaction, motivation, and
productivity (Kluger and DeNisi 1996). Consider, for
instance, a setting in which workers are compen-
sated using piece rates, so that better performers also
receive higher compensation. If workers display rela-
tive concerns (Festinger 1954, Frank 1985, Buunk and
Gibbons 2007), information on relative performance
will affect well-being and effort decisions even when
pay is independent of relative performance per se.
Understanding this direct motivation effect is
important for several reasons. First, at a general level
it can contribute to our knowledge of the interlinks
between preferences, behavior, and the information
available to individuals. Second, it can shed light
on the additional, probably unintended, effects of
explicit incentive schemes, like tournaments, that rely
on comparisons across workers. Third, it has direct
implications for optimal communication policies. In
organizations where information on relative standing
can be computed easily, there is clearly potential for
a judicious communication policy to increase produc-
tivity at little or no cost to the firm.
In this paper we use a field quasi experiment to
study the direct effect of informing workers about
their own position in the productivity and pay dis-
tribution. There is a scarcity of evidence using real
workers in real economic conditions to study the
effects of relative performance feedback. In addition
to the usual difficulties in obtaining data from orga-
nizations, there is the additional difficulty that, when
relative performance is communicated in firms, it typ-
ically has explicit or implicit monetary consequences.
For instance, a worker being informed that he has
performed better than his colleagues may reasonably
conclude that he has a chance of receiving a promo-
tion to a better paid position, which could translate
into higher effort. This makes it difficult to disentan-
gle the direct effect of relative performance feedback
from the effect of the monetary reward to which such
feedback is often linked.
In this paper we take advantage of a unique
microlevel data set, constructed from a German
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1722 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
wholesale and retail organization. In the main ware-
house of this organization, workers were being paid
piece rates. From some point onward, however, they
began to receive private information on their own
position in the distribution of pay and productivity.
For reasons that we describe in detail below, this rel-
ative performance information exercise (henceforth,
RPIE) did not have (and was known to not have)
any type of compensation or career consequences. Its
introduction was triggered by exogenous reasons and
was not part of any wider managerial policy.
Although not a randomized trial, a key feature of
our data set, namely, the availability of information
on individual daily performance, provides the basis
for a quasi-experimental research design. We make
use of this rich information to estimate the produc-
tivity increase following the introduction of the RPIE.
We find the immediate average increase to be 6.8%
and find no evidence that this effect vanishes over
time. The increase is consistent across the distribu-
tion of productivity. We also show that this long-term
increase in productivity (i.e., in “quantity”) did not
occur at the expense of either a long-term decrease in
the quality of the work produced or an increase in the
quit rate among the workforce.
To understand the economic mechanism behind
this increase in productivity, we exploit the differen-
tial timing in the two parts that constitute the RPIE.
The information was communicated on the first day
of the month and referred to the relative perfor-
mance during the month before the previous one. For
instance, on November 1, workers learned about their
performance during the month of September. As a
result of this timing, we can distinguish the “kick-
off moment” (the date in which workers performance
first started to count toward the information that they
would receive) from the “revelation moment” (the
date in which workers first learned their relative per-
formance). We find that productivity increases not
only at the revelation moment, but also at the kickoff
moment. The fact that productivity starts increasing
even before workers actually learn any new informa-
tion casts doubt on most potential explanations of our
findings, including those based on pure worker learn-
ing about absolute ability or the distribution of wages.
Our preferred interpretation suggests that the
observed increase in productivity was caused by an
increase in concerns about relative standing. As we
discuss in detail below, our results are consistent with
information about relative performance and the antic-
ipation of such information leading to an increase in
relative concerns that shifted the distribution of pro-
ductivity to the right, because all types of workers
tried to improve their relative position in terms of
pay and/or productivity. Our findings therefore add
to previous studies arguing that workers are affected
by relative concerns, whereby increases in others’ per-
formance always has a negative effect on own utility
(Luttmer 2005, Charness and Kuhn 2007, Clark et al.
2008). It has been argued that the existence of such
concerns leads to excessive levels of effort and con-
sumption, creating a role for welfare increasing public
policy (Oswald 1997). We interpret the findings of our
paper as providing evidence on the existence of “con-
tests for relative position” using field data.
1.1. Related Literature
Psychologists and management scholars have long
recognized that workers are motivated by more than
monetary rewards (Etzioni 1971, Deci 1975). Individ-
uals exert effort at work partly because of interest
or enjoyment in the task itself or because completing
the task provides a feeling of accomplishment or self-
worth. Typically, the presence of extrinsic rewards is
believed to have the effect of “crowding out” intrin-
sic motivation (Kohn 1993, Deci et al. 1999). At the
same time, Deci (1972) and Anderson et al. (1976) find
that (positive) feedback increases intrinsic motivation.
In our setting, we argue that relative performance
feedback does indeed increase relative concerns and
therefore intrinsic motivation. The existence of perfor-
mance pay most likely reinforced the strengthening
of relative concerns following the RPIE and therefore
served to enhance the increase in intrinsic motivation.
Feedback interventions can sometimes lead to a
decrease in performance (Kluger and DeNisi 1996).
A substantial body of work in human resource
management and applied psychology has therefore
analyzed the circumstances that make job feedback
enhance satisfaction and performance (Nadler 1979,
Latham and Wexley 1981, Moore and Klein 2008).
Because in most organizations evaluations are mainly
subjective, research efforts have often focused on
which organizational actors should contribute to the
evaluation, what format the rating should take, and
how to best present it to the evaluated worker (for an
overview, see, for instance, London 2003). In our firm,
performance was easily quantifiable, and therefore
it was possible to be both objective and detailed in
the information presented to workers.
the objective nature of performance and the clarity
of the goal implied that workers dissatisfied with
their relative performance could easily understand
that a reduction in the gap between expectations and
feedback was possible through the exertion of higher
“Constructive” feedback is regarded as more likely to lead to
increased performance (Baron 1988). London (2003, pp. 15–16)
argues that feedback is constructive when “it offers concrete infor-
mation that can be used,” “it is clear and easily understood,” and
“it is interpreted similarly by the source and recipient.” All these
characteristics clearly apply to the RPIE that we study.
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
Management Science 57(10), pp. 1721–1736, ©2011 INFORMS 1723
effort (Kluger and DeNisi 1996, Bandura and Cervone
1983). This fact may contribute to explain the positive
effect on productivity that we document.
In economics, Eriksson et al. (2009) find no effect
of providing relative performance feedback to labo-
ratory subjects compensated via piece rates. Kuhnen
and Tymula (2010) develop a theoretical model of self-
esteem, study subjects in a laboratory flat wage setting,
and find that relative performance feedback has both
ex ante (i.e., for agents anticipating it) effects and ex
post (for agents receiving it) effort effects. Their result
echoes our finding that the RPIE affects performance
both at the “kickoff stage” andat the “revelation stage.”
Two recent experimental papers try to separately
measure the importance of cardinal versus ordi-
nal considerations. Clark et al. (2010) find that
agents’ rank in the income distribution affects effort,
although, controlling for rank, relative income does
not. Sunde et al. (2011) observe laboratory subjects
using functional magnetic resonance imaging. They
find that both relative income and rank affect regions
of the brain associated with the processing of basic
rewards. In our firm, the RPIE consisted of both cardi-
nal and ordinal information, and unfortunately their
relative importance cannot be separately identified.
Several papers study relative performance feedback
using field data. In a context where workers are com-
pensated according to team performance, Delfgaauw
et al. (2009) find that information on relative perfor-
mance leads to higher productivity, whereas Bandiera
et al. (2009) find the opposite. In a schooling context,
Azmat and Iriberri (2010) and Tran and Zeckhauser
(2009) argue that learning about relative performance
leads to higher student effort.
Finally, this paper contributes to the literature on
tournament schemes by documenting a new mar-
gin through which competition operates, namely, the
“pure contest for relative position,” which arises
when individuals derive utility directly from out-
performing others in terms of performance or pay.
Importantly, incentives along this margin are free to
the principal: the 6.8% average increase in produc-
tivity that we observe in our setting was achieved
at no cost to the firm, other than the negligible cost
of communicating with workers. Thus, even tourna-
ments without prizes can have large incentive effects.
1.2. Outline of This Paper
Section 2 discusses the institutional setting in which
the RPIE occured. Section 3 presents and interprets
our main results. In §4 we examine other effects of
the RPIE. Section 5 concludes.
2. Institutional Setting
2.1. Workplace Conditions and Compensation
We study an intervention occuring in the second half
of 2001 in the main warehouse of a German wholesale
and retail organization. The company’s wholesale
division handles a large number of food and nonfood
products, which are sold to independent supermar-
kets as well as the company’s own retail division.
During our sample period, the firm employed
around 65 workers to perform the core task of the
warehouse. This task involves picking up customer
orders, assembling the requested products, packing
them onto a trolley, and moving the trolley to the
goods-out area of the warehouse.
Work is strictly
individual, and opportunities to observe each other’s
work are limited. Worker’s salary comes in three
parts: a fixed base salary, which is the same for
every worker,
a “productivity” or “quantity” perfor-
mance component, and a “quality” performance com-
ponent. The structure of the compensation scheme
was constant between August 1999 and August 2002.
The levels of the piece rates have, however, changed
over time.
The “quantity” performance component has two
parts, with their corresponding linear piece rates: the
first part depends on the number of orders completed,
and the second on the number of goods dispatched.
The number of goods dispatched is obviously the
product of the number of orders and the average
number of goods per order. In 2001 the “quantity”
performance component represented around 25% of
the average worker total compensation.
The “quality” performance component depends on
the number of wrongly dispatched goods and, in
practice, is much less important. When a customer is
sent the wrong good and he complains to the firm,
this fact is recorded and tracked back to the individ-
ual worker in charge of that particular good. A small
linear discount is applied to a worker’s total compen-
sation when his monthly mistake rate exceeds a cer-
tain threshold. In 2001 the average discount was less
than 1% of total compensation.
Both performance components are paid with a one
month delay. Hence, in the pay slip received, say,
on November 1, workers learn about performance
and wage per hour corresponding to the month of
September, rather than October.
Our measure of productivity is the average number
of goods dispatched per hour by a given worker on a
given day. Because customer orders contain a varying
number of goods, we always control for the average
The allocation of orders to workers took place following a “cab
rank system,” so that a worker finishing an order would be
matched with the order at the top of the list, and so on. In regres-
sions available upon request we found that a worker’s productivity
in a month could not predict the characteristics of the orders that,
on average, he would receive the following month.
In this firm there is a one-to-one mapping between performance
and pay. For this reason, we will not be able to disentangle whether
concerns about relative standing refer to performance or to pay.
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1724 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
number of goods per order. Otherwise we would be
overestimating the productivity of those workers that
handled a few large orders rather than many small
ones. Once the number of goods per order is held
constant, the number of goods per hour and the num-
ber of orders per hour are equivalent measures of
Upon joining the firm, workers face a six month
probationary period, and they may not be offered
a regular contract at the end of this period. How-
ever, once they survive the probationary period they
are very unlikely to be fired. After two years, the
restrictions imposed by German law, the fact that less
productive workers are cheaper to the firm, and the
screening performed during the probationary period
make termination of contracts all but unheard of.
If the job that we study is unusually secure, it is
also a cul-de-sac, because the skills acquired in it
do not transfer easily to other positions in the firm.
Among the 207 workers holding this job during the
last 10 years, only 2 workers have ever been pro-
moted. The reason for this is that there is no natural
higher-level position for which the skills gained in
this job provide either good training or a good signal
of future productivity. Thus, any form of career con-
cern is a truly unlikely source of motivation for the
workers in our study.
Last, there is a certain amount of learning and skill
in this occupation. This is because being able to iden-
tify the optimal route to gather the goods ordered
by a customer leads to a significant gain in time
and productivity. This ability/skill component trans-
lates into a strong serial correlation in workers’ rel-
ative performance across time. During our sample
period, workers’ rank-order position in the productiv-
ity distribution during a certain month had a correla-
tion of 0.88 with their position in the previous month.
2.2. The Relative Performance
Information Exercise
In the summer of 2001, a few members of the
workforce—employees are not unionized—requested
from management access to information about the
wage per hour being earned by the average worker.
Given the low computational cost of creating and
distributing this information, the firm agreed to
this demand. All workers were then notified that
their September performance report—that is, the one
included with the November 1, 2001, pay slip—would
be the first one for which they would receive this
In addition, management decided to also commu-
nicate to workers their own rank-order position in
the productivity/wage per hour distribution. The rea-
son for this was very specific: two workers had been
constantly complaining about the conditions of the
job and spreading discontent among the workforce.
These two workers happened to be among the worst
performers, so management thought that privately
revealing to them this fact would help to alter their
The introduction of the RPIE provides the basis
of a quasi-experimental research design. Note that
the RPIE was first triggered by the demand of a
small number of workers and then by a very spe-
cific managerial motivation that was orthogonal to
the compensation scheme in place. Furthermore, com-
pany insiders stated to us that the introduction of
the RPIE did not coincide with any type of shock to
the dispatching technology. The identification strat-
egy, which we discuss further below, uses both indi-
vidual daily productivity and a short time window
around the introduction of the RPIE to ensure that
potential preexisting productivity trends are not con-
founding our empirical findings.
The information finally communicated on Novem-
ber 1 and subsequent months consisted of the mini-
mum, maximum, and average wage per hour and a
worker’s rank-order position in the wage per hour
distribution. Workers were notified prior to Septem-
ber 1 that they would receive this information.
3. The Productivity Effects
of the Relative Performance
Information Exercise
Our discussion above suggests two moments at which
workers’ behavior may have changed as a result
of the RPIE. First, on September 1—the “kickoff
moment”—workers’ daily productivity first started to
count toward the RPIE, and they may have altered
their effort in response to this fact. If workers antici-
pate that they will derive utility tomorrow from learn-
ing that they outperformed their colleagues today,
then increasing effort today to obtain that future
utility seems like a reasonable response. Second,
on November 1—the “revelation moment”—workers
received information about their relative performance
for the first time. This may have led to a further
behavioral response if receiving such information fur-
ther increased the salience and importance of relative
3.1. Descriptive Analysis
We first provide graphical evidence that the RPIE was
associated with a large and sudden shift in workers’
According to company sources, these two workers were indeed
embarrased to learn that they ranked at the bottom of the distribu-
tion and stopped complaining. Their contracts were not terminated
nor were they disciplined or even threatened in any way, a fact that
should alleviate any lingering concerns that workers’ reaction may
have been due to the fear of job terminations.
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
Management Science 57(10), pp. 1721–1736, ©2011 INFORMS 1725
Figure 1 Evolution of Productivity and Other Variables Over Time
Revelation Kickoff

Aug99 Sep01 Nov01 Aug02
Aug99 Sep01 Nov01 Aug02
Panel C
Revelation Kickoff
Aug99 Sep01 Nov01 Aug02
Aug99 Sep01 Nov01 Aug02
Panel A


Panel B
Panel D
Kickoff Revelation
Notes. Dependent variables are the number of goods dispatched per hour worked (panel A), the ratio of fresh goods over fresh and dry goods (panel B),
the average value of a good (panel C), and worker experience (panel D). Each dot is the (averaged across workers) dependent variable of a different working
day. Vertical dashed lines display September 1, 2001 (kickoff date), and November 1, 2001 (revelation date). In panels A–C, solid lines display local linear
regressions (with an eight-day bandwidth) computed separately before the kickoff date, after the revelation date, and in between.
productivity and not with similar discontinuities in
observable determinants of productivity.
Figure 1A shows the evolution of daily productiv-
ity averaged across workers. It is clear from the fig-
ure that the introduction of the RPIE coincided with a
sharp increase in productivity. Before September 2001,
an average worker dispatched less than 145 goods per
hour, whereas by November 2001 productivity had
increased to around 157 goods per hour. Such increase
seems out of line both with the flat or mildly pos-
itive preexisting trend and with preexisting fluctua-
tions around the trend.
In Figure 1B and C, we display the evolution of
characteristics of the goods dispatched by workers.
Goods can be either fresh or dry, and it is claimed
by company insiders that fresh goods take longer to
dispatch because the aisles where they are placed are
narrower, and this creates jams in the warehouse. In
early 2001, a weekend shift was added, leading to a
more extreme concentration of fresh and dry goods in
different days of the week. However, it is clear from
Figure 1B that the RPIE did not coincide with any big
discontinuity in the ratio of fresh goods. Similarly, we
plot the evolution of the average value of the goods
dispatched by workers over time in Figure 1C. Again,
no large and permanent discontinuity seems apparent
from the figure.
In Figure 1D we display the evolution of aver-
age worker experience over time. Average experience
trends upward if there are no changes to person-
nel and instead decreases in periods, like the second
part of 2000, in which the firm hires new workers.
Such fluctuations are obvious in the figure, but, again,
no large changes are apparent during our period of
It seems difficult from Figure 1A to attribute the
sharp increase during the September 2001–November
2001 period to any preexisting trend in productivity.
To ensure, however, that potential mispecification
of preexisting time trends does not confound our
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1726 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
Table 1 Descriptive Statistics
(1) (2) (3) (4)
Mean SD Minimum Maximum
Panel A: Firm-level characteristics
Goods per day 1,107.24 152.73 474.78 1,561.79
Goods per hour 149.73 7.85 126.84 172.14
Average value of good 100 28.57 3.6 122.22
Ratio of fresh goods 0.56 0.22 0.22 1
Number of observations 152
Panel B: Individual-level characteristics
Male 0.63 0.48 0 1
Experience (in years) 3.99 4.06 0.50 23.88
Age (in years) 38.75 8.35 21.27 58.48
Left December 31, 2007? 0.38 0.49 0 1
Ranking August 2001 34.01 18.35 3 66
Number of observations 63
Notes. Descriptive statistics are based on the sample comprising July 1,
2001, to December 28, 2001. The number of worker-day observations
included in the sample is 5,275. Panel A displays variables at firm-day level
by average worker. Average value of a good is normalized at 100. Panel B dis-
plays variables at the worker level. The sample is restricted to those workers
with at least six months of experience in the firm.
estimates, we focus our main analysis on a short time
window comprising six months in total. Our main
sample consists of three periods: (1) the “baseline
period” includes the two months (July and August)
just before the introduction of the RPIE; (2) the
“kickoff period” includes the two months (Septem-
ber and October) during which workers did not
have any new information but knew that their daily
productivity would form the basis on which relative
performance would be calculated; and (3) the “rev-
elation period” includes the two months (November
and December) right after workers first learned their
relative performance.
Using this sample, panel A of Table 1 displays
the main input and output variables at firm level. In
panel B of Table 1 we display a number of worker
characteristics. The average worker in our sample has
been at this job for four years and is 39 years of age.
More than 60% of workers in our sample were still at
the job six and a half years later, which implies that
the yearly separation rate was just 6%. We also dis-
play information on the rank-order performance dur-
ing August 2001, which has been constructed by us
but was not computed by the firm.
We restrict our study to workers with more than six months of
experience. We do this for several reasons. First, the returns to expe-
rience are both very steep and very concave among workers in their
first six months. As a result, controlling for a linear trend may not
adequately capture the evolution of productivity for these workers,
even in such a short time horizon as the one used here. Second,
these workers are in their probationary period and may have dif-
ferent incentives from those faced by the rest of the workforce. In
particular, we cannot rule out the possibility that they may perceive
3.2. Econometric Framework
Our basic estimating equation takes the following lin-
ear form:

· o +j
+g(|) +e
, (1)
where j
is (the log of) worker i’s productivity, mea-
sured in number of goods dispatched per hour on day
|; K
= 1 in the post–September 1 period; and í
= 1
in the post–November 1 period. Our model also con-
tains a vector of covariates X

, including, as discussed
above, the (log of the) average number of goods per
customer order by worker i on day |. We allow for
individual time-invariant unobserved effects, j
, to
account for the possibility that individual workers
may differ across time periods.
Our identification strategy relies on the comparison
of the same workers over time. Relying on such varia-
tion implies that we cannot be completely nonrestric-
tive in accounting for time effects. We instead allow
productivity to evolve smoothly over time through
the parametric function g(|). We show however that
our results are robust to the specification of g(|) as
well as to the period over which g(|) is calculated. The
variable e
captures independent and identically dis-
tributed (iid) person-specific idiosyncratic shocks.
To account for correlation of the observations from
the same date, we adjust the standard errors by allow-
ing for arbitrary variance–covariance matrix within
each date across individuals. In practice, this does not
have much effect on the standard errors.
Our parameters of interest are p
and p
. The iden-
tifying assumption in estimating these parameters
from Equation (1) is that, without the introduction of
the RPIE, productivity would not have deviated from
the trend g(|).
3.3. Baseline Results
Table 2 displays our baseline results. Column (1)
shows that productivity was 2.7% higher in the kick-
off period and 5.9% higher in the revelation period,
relative to the baseline period.
The fit of the model is very large even in this first
specification with just three explanatory variables.
The reason is that the number of goods per order is a
very strong determinant of the number of goods dis-
patched per hour, because workers can dispatch more
goods when they do not have to make constant trips
to the goods-out area of the warehouse and to pick
up new orders. This fact underlines the importance of
controlling for this variable in every specification.
the information on their relative position as a signal on the likeli-
hood that their contract will be renewed. By focusing on workers
unlikely to be making those considerations, we will be able to argue
that pure concerns for relative position lie behind their reaction to
the RPIE. This restriction reduces our sample by just three workers.
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
Management Science 57(10), pp. 1721–1736, ©2011 INFORMS 1727
Table 2 Baseline Results
(1) (2) (3)
Unconditional Time trend Other covariates
Kickoff 0.027
(0.005) (0.010) (0.006)
Revelation 0.032
(0.006) (0.010) (0.007)
(log)Goods per order 0.308
(0.011) (0.011) (0.006)
Time trend −0.0001 −0.0001

(0.0001) (0.00007)
(log)Goods piece rate −0.045
Ratio fresh goods 0.013
Average value of goods −0.174
(in euros) (0.022)
Day of week dummies No No Yes
Worker experience dummies No No Yes
Worker fixed effects No No Yes
Adjusted l
0.22 0.22 0.77
Number of observations 5,275 5,275 5,275
Notes. Dependent variable is (log of) number of goods dispatched per hour
worked. DatacompriseJuly1, 2001, toDecember 28, 2001. Kickoff isadummy
variable taking value 1 in the post–September 1, 2001, period. Revelation is a
dummy variable taking value 1 in the post–November 1, 2001, period. Worker
experience dummies capture the quarter of experience of the worker. Sample
is restricted to those workers with at least six months of experience in the firm.
Robust standard errors allowfor cluster at the day level.

, and
denote estimates significant at 10%, 5%, and 1%,
Column (2) adds a linear time trend to the specifi-
cation. The estimated coefficients of interest are very
similar in magnitude to those of column (1), and
the estimated linear trend is not statistically different
from zero. Thus, unobserved time-varying factors cor-
related with productivity are unlikely to be biasing
our estimates.
In column (3) we add extra covariates, day of week
dummies, experience dummies, and worker fixed
effects. We find the average effect of kickstarting the
RPIE to be 2.8%, whereas the average effect of actu-
ally revealing the relative performance information is
4.0%. The cumulative effect is 6.8%.
The coefficients on the covariates are interesting in
their own right. Consider the level of the goods piece
rate, which the firm altered in every month of our
sample. These changes were by very small percent-
ages (usually 1% or 2%), so it is plausible that the
workforce barely (if at all) reacted to them. The neg-
ative and statistically insignificant coefficient on the
piece rate indicates indeed that workers did not seem
to work harder in months in which the piece rate was
marginally higher.
We find, however, that workers were more pro-
ductive when the goods dispatched were cheaper.
Company insiders suggested that this may be either
because cheaper goods are smaller in size or because
they are more common and workers are better aware
of their location.
3.4. Placebo RPIE and Robustness Checks
Because the RPIE was introduced at the same time for
all workers, identification of its effect on productiv-
ity arises from a comparison over time of the same
worker. The estimated effects are therefore biased
upward to the extent that they capture factors that
cause productivity to rise independently of the RPIE
and that are not captured by the linear trend. Table 3
presents a number of robustness checks to address
this and other concerns.
To judge whether our findings are due to season-
ality, we repeat the estimation of (1) substituting our
baseline sample from 2001 with an equivalent sample
from the previous year. This new “placebo” sample
spans July 1, 2000, to December 31, 2000, and con-
sists of 3,836 observations. We define a placebo kickoff
dummy to be equal to one after September 1, 2000,
and a placebo revelation dummy to be equal to one after
November 1, 2000. We include worker fixed effects, a
time trend, experience dummies, and the other con-
trols. Column (2) of Table 3 suggests that our main
results were not due to seasonality. The estimated
coefficients are much smaller in magnitude and are
either not statistically significant or weakly significant
but with a negative sign.
We next examine the robustness of our estimates to
the sample width by estimating (1) on a sample com-
prising 12 months in total. As column (3) of Table 3
shows, the estimated coefficients are still statistically
significant and only slightly different in magnitude.
Second, we study whether our results are robust
to allowing the parametric time trend g(|) to take
different formats. We first allow g(|) to take a
quadratic form. Because a problem of severe collinear-
ity between | and |
arises in our baseline time win-
dow of 6 months, we use a window of 12 months.
Column (4) of Table 3 shows that our main findings
are indeed robust to allowing g(|) to take a more flex-
ible form.
In our benchmark specification from column (3) in
Table 2, the linear trend is estimated using both pre-
treatment and treatment observations. Alternatively,
we may want to estimate the trend exclusively using
pretreatment information. This allows us to evaluate
the effect of the RPIE above or below what we would
expect productivity to be given its pretreatment trend.
To do this, we estimate the empirical model

· o +j
+0(1 −K
) · | +e
and use the estimated coefficients to project the
expected value of productivity ˆ j
given the estimated
pretreatment trend
0. We then compute the difference
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1728 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
Table 3 Placebo RPIE and Robustness Checks
(1) (2) (3) (4) (5) (6)
Baseline 12 months Quadratic Pretreatment Pretreatment
results Placebo window trend trend effects
Kickoff 0.028
(0.006) (0.005) (0.006) (0.005) (0.005)
Placebo kickoff 0.008
Revelation 0.040
(0.006) (0.003) (0.004) (0.005) (0.006)
Placebo revelation −0.011

Table 2 covariates Yes Yes Yes Yes Yes Yes
Standard errors Day Day Day Day Boot- Boot-
cluster cluster cluster cluster strapped strapped
Adjusted l
0.74 0.71 0.74 0.74 0.74 0.74
Window months 6 6 12 12 6 6
Observations 5,275 3,836 10,176 10,176 5,275 5,275
Notes. Dependent variable is (log of) number of goods dispatched per hour worked. Column (1) is baseline regression
(i.e., column (3) from Table 2). Column (2) data comprise July 1, 2000, to December 31, 2000, and the placebo kickoff
and placebo revelation dummies take a value of 1 in the post–September 1, 2000, and post–November 1, 2000, periods,
respectively. In columns (3) and (4) data comprise April 1, 2001, to March 31, 2002. Column (4) includes a quadratic
time trend. In column (5), the linear time trend is estimated exclusively on the baseline period (July and August 2001).
In column (6), the linear time trend, the worker fixed effects, and the effects of the covariates are estimated exclusively
on the baseline period (July and August 2001). Robust standard errors allow for cluster at the day level unless specified

, and
denote estimates significant at 10%, 5%, and 1%, respectively.
between actual and predicted productivity ˆ u
= j

ˆ j
and estimate
ˆ u
. (3)
The estimated coefficients
are displayed in
column (5) of Table 3. The coefficients are very similar
to those of our benchmark specification from Table 2,
suggesting that our results are robust to the form of
the parametric time trend.
In column (6) we use only pretreatment informa-
tion to estimate the individual fixed effects, j
, and
the effects of other covariates, o, as well as the linear
trend, 0. Again, we find that the estimated coefficients
are statistically significant and similar in magnitude.
3.5. Heterogeneous Effects
Table 2 shows that the average effect of the RPIE is
6.8%. The effect may however be different for differ-
ent types of workers, perhaps even negative for some.
One important limitation of our study is the absence of a con-
trol group. To partially remedy this drawback, we examined the
performance of a comparable German food trader, Metro Group.
Productivity measures at the warehouse worker level were unfor-
tunately unavailable. Instead, we studied the evolution over time
of a quarterly measure of operating profits, earnings before interest,
taxes, depreciation, and amortization. We found that this measure
of performance was not characterized by any significant disconti-
nuity in the second part of 2001 (details are available upon request
from the authors).
In Table 4 we split the workforce on the basis
of observable characteristics and run our benchmark
specification (1) separately for different groups of
workers. Panel A of Table 4 shows that the reactions
of males and females to the RPIE are very similar.
Although the estimated coefficients are slightly differ-
ent in magnitude, these differences are not statistically
significant. In panel B we find that only workers in
the middle and top thirds in terms of experience react
to the kickoff of the RPIE. On the other hand, we find
no differences in the revelation coefficient for workers
with different levels of experience. In panel C we use
the (undisclosed) ranking from August 2001 to split
the sample by preexisting level of performance. We
find weak evidence that workers in the bottom third
increase their productivity more at the kickoff stage,
whereas workers in the top third increase in produc-
tivity more at the revelation stage. The overall effect
of the RPIE is very similar for workers with different
preexisting levels of performance.
Instead of aggregating workers on the basis of
their observable characteristics, we can study sepa-
rately the reaction of every individual worker to the
RPIE. To do this, we ran 58 separate worker-level
regressions equivalent to (1) and computed the sums
of the estimated kickoff and revelation coefficients.
Although there was substantial heterogeneity across
workers, the reaction to the RPIE was positive for
the overwhelming majority of workers (i.e., 51 of 58).
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
Management Science 57(10), pp. 1721–1736, ©2011 INFORMS 1729
Table 4 Decomposition by Gender Experience and Performance (August Ranking)
Panel A: Decomposition by gender
(1) (2) (1)–(2)
Males Females p-value
Kickoff 0.023
(0.008) (0.011)
Revelation 0.047
(0.008) (0.012)
Worker FE Yes Yes
Covariates Yes Yes
Observations 3,307 1,968
Panel B: Decomposition by experience
(1) (2) (3) (1)–(2) (1)–(3) (2)–(3)
Least experience Middle experience Most experience p-value p-value p-value
Kickoff −0.010 0.029
0.02 0.00 0.41
(0.012) (0.010) (0.010)
Revelation 0.045
0.98 0.55 0.56
(0.013) (0.012) (0.011)
Worker FE Yes Yes Yes
Covariates Yes Yes Yes
Observations 1,484 1,833 1,907
Panel C: Decomposition by performance (August ranking)
(1) (2) (3) (1)–(2) (1)–(3) (2)–(3)
Bottom ranked Medium ranked Top ranked p-value p-value p-value
Kickoff 0.052
0.015 0.019

0.02 0.04 0.78
(0.012) (0.009) (0.011)
Revelation 0.035
0.72 0.25 0.11
(0.014) (0.011) (0.011)
Worker FE Yes Yes Yes
Covariates Yes Yes Yes
Observations 1,553 1,860 1,862
Notes. The dependent variable is (log of) number of goods produced per hour worked. Data comprise July 1, 2001, to December 28,
2001. The sample is restricted to those workers with at least six months of experience in the firm. Covariates are as in column (3)
of Table 2, except in panel B, which excludes experience quarter dummies. Robust standard errors allow for cluster at the day level.
FE, fixed effects.

, and
denote estimates significant at 10%, 5%, and 1%, respectively.
Furthermore, of the 19 workers for whom the com-
puted overall effect was statistically different from
zero at the 10% level, 18 had a positive sign, whereas
only 1 had a negative sign.
We argue in §3.7 that two findings from this sec-
tion are unlikely to be compatible with basic models
of learning about ability or the distribution of perfor-
mance: (a) very experienced workers do not react less
at the revelation stage than less experienced workers,
and (b) the effect of the RPIE can be regarded as pos-
itive for almost all workers.
3.6. Effects on Inequality
We now study whether the RPIE affected the amount
of inequality across workers. Figure 2 plots the kernel
distribution of productivity for August, September,
and November 2001. It seems that the RPIE shifted the
distribution to the right, without noticeable effects on
the dispersion of productivity. To confirm this finding
formally, we computed the coefficient of variation for
the productivity variable for each day of our baseline
sample. We found that this measure of dispersion was
not different on average following the introduction of
the RPIE (details are available upon request from the
authors). Thus, static inequality did not change.
We can also study inequality from a dynamic per-
spective by examining the likelihood that a worker at
a position in the performance distribution in a partic-
ular month repeats the position the following month.
In Figure 3 we plot workers’ rank position in terms
of wage per hour across consecutive months for each
month in our baseline sample. There is significant per-
Overall, it does not seem as if persistence
We do not find the existence of strong serial correlation in rank
incompatible with the notion that agents exert more effort when
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1730 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
Figure 2 Distribution of Productivity Over Time
200 150
August 2001 (baseline)
September 2001 (kickoff)
November 2001 (revelation)
Productivity (number of goods per hour)
Notes. This figure displays the kernel densities of the productivity variable
for three selected months: August 2001 (the last baseline month), Septem-
ber 2001 (the first kickoff month), and November 2001 (the first revelation
month). Productivity is defined in terms of goods dispatched per hour.
in rank position varied following the introduction of
the RPIE. We confirmed this by regressing a worker’s
rank position in a month on his position in the pre-
vious month and interacting this variable with the
kickoff and revelation variables. We found that the
coefficients associated with these interactions were
not statistically different from zero (details are avail-
able upon request from the authors).
To summarize, the RPIE led to an increase in aver-
age productivity and wages while leaving both static
and dynamic disparities between workers essentially
3.7. Interpretation and Alternative Explanations
To summarize the main findings: (a) the effect of the
RPIE is on average positive both at the kickoff stage
and at the revelation stage (Figure 1A and Table 2);
(b) individual regressions show that the effect is pos-
itive for a large proportion of workers and negative
for just one worker (§3.5); (c) the increase in produc-
tivity at the kickoff stage is larger for workers with
more experience (panel B of Table 4) and for workers
at the bottom of the performance distribution (panel
C of Table 4); and (d) the overall increase in produc-
tivity is, however, statistically equal for workers with
different observable characteristics (Table 4).
We now discuss the potential explanations of our
their relative concerns have stronger. This is because (a) relative
concerns could be cardinal as well as ordinal, (b) rank mobility
over time can be high in the presence of strong serial correlation
provided that the number of workers is large, and (c) increased
relative concerns could be affecting the marginal return to effort
even while having no effect on equilibrium rank positions.
3.7.1. Preferred Interpretation: Increase in Rel-
ative Concerns. We regard the evidence presented
here as consistent with the notion that workers’ incip-
ient concerns about relative standing were activated
by the RPIE. This interpretation is consistent with
the documented productivity increase at the kickoff
stage. By increasing effort at that stage, a worker
would have increased his chances of outperforming
his colleagues, a fact that, on being learned later,
would generate a utility increase. This explanation is
also consistent with the revelation of relative perfor-
mance further increasing the salience of these rela-
tive concerns and leading to the additional increase
in productivity documented in Table 2. Last, the exis-
tence of concerns for relative pay is perfectly com-
patible with more experienced workers reacting more
strongly at the kickoff stage, because it is natural to
think that these concerns should be stronger for indi-
viduals that have been members of the workforce for
a longer period of time (Frank 1985).
What type of relative concerns can explain our find-
ings? Consider first relative concerns of the “inequity
aversion” type (Fehr and Schmidt 1999, Fehr and
Gächter 2000). These preferences would be consistent
with workers at the bottom of the distribution increas-
ing their effort by a large amount, and with workers
at the top not working harder and even decreasing
their effort to allow the others to catch up. Yet, this is
the opposite of what we find in §3.5.
We instead regard relative concerns whereby in-
creases in others’ performance always have a nega-
tive effect on own utility (Frank 1984, Fershtman et al.
2005, Moldovanu et al. 2007) as consistent with our
findings. As Frank (1984), Fershtman et al. (2005), and
Moldovanu et al. (2007) discuss, an increase in these
concerns translates into everybody increasing their
level of effort and nobody decreasing it, as we find
in §3.5.
We now discuss alternative explanations.
3.7.2. Alternative Explanation: Career Concerns.
Could concerns about either termination of employ-
ment or a potential promotion present an explanation
of our findings? As discussed in §2, the likelihood
of both terminations and promotions in our firm is
close to zero. Working conditions and base pay are
fixed and identical for every worker, regardless of per-
formance. Furthermore, workers were explicitly told
that the relative performance information would not
be used in any way by management. Thus, we do not
regard career concerns as a valid explanation.
Using career concern interpretations, it is also diffi-
cult to explain why productivity did not decrease for
almost any worker in the firm. Consider for instance
the incentives of a weak performer who, before the
introduction of the RPIE, is unaware of this fact
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
Management Science 57(10), pp. 1721–1736, ©2011 INFORMS 1731
Figure 3 Relation Between Ranks in Two Consecutive Months
0 0.5 1.0 0 0.5 1.0 0 0.5 1.0
Baseline 1 (July 2001) Baseline 2 (August 2001) Kickoff 1 (September 2001)
Kickoff 2 (October 2001) Revelation 1 (November 2001) Revelation 2 (December 2001)



Normalized rank previous month
Notes. This figure presents a scatter plot capturing the relation between the rank (in terms of the wage per hour) obtained by an individual in a month and the
rank obtained by the same individual in the previous month. The ranks are all normalized to one by dividing by the total number of workers working in the firm
on that month. Six different scatter plots are presented for each of the months comprising our baseline sample.
and exerts some extra effort with the hope of rank-
ing among the top and meriting a promotion. We
would expect that upon learning his mediocrity, he
would regard a promotion as virtually impossible and
decrease his efforts to attain it. Yet we do not find this
to be the case for almost any worker. Similarly, very
able workers fearing a potential termination should
have put less effort, not more, after learning that they
ranked at the top of the distribution. Again, this is the
opposite of what we find.
3.7.3. Alternative Explanation: Learning About
Own Absolute Ability. One potential explanation of
our findings is that the RPIE may have helped work-
ers to learn about their own ability for the job, and this
may have led to a readjustment in the chosen level of
effort. One potential benchmark for this argument is
provided by Ertac (2006), who presents a theoretical
model in which effort and ability are strategic comple-
ments in a worker’s production function, and workers
are uncertain about their own ability. Past production
can help workers learn about their own ability but is
subject to an unobserved firmwide productivity shock
that makes learning imperfect. In this setting, Ertac
(2006) shows that information on how much other
workers produced can help a worker to learn about
their own ability. For instance, a worker with low pro-
ductivity in a month will revise expectations about his
own ability upward after observing that others also
did badly, because his low productivity is likely to
be the result of an unfavorable firmwide productivity
shock rather than the result of low ability.
We see two problems with this explanation.
The main problem is that any interpretation based
exclusively on learning cannot explain the existence
of a positive productivity effect at the kickoff stage,
in which workers do not learn anything.
An additional problem is that the predictions of the
model outlined above do not seem consistent with
our empirical findings. In the model outlined above
the arrival of new information makes some workers
revise their expected ability upward and exert more
effort, whereas other workers revise it downward and
exert less effort.
This is not what we found in the
data. We found in §3.5 that there is almost no hetero-
geneity in the sign of the reaction to the RPIE: almost
all workers react by increasing their effort.
To gain intuition, consider a situation in which all workers have the
same prior expected ability and learn the average level of produc-
tion of their coworkers. Roughly speaking, workers producing more
(respectively, less) than the average will update their ability upward
(downward) and exert more (less) effort upon learning this fact.
A second prediction of Ertac (2006) is that workers less uncertain
about their own ability should react less to the information con-
veyed by the RPIE. Again, this is inconsistent with our findings.
We find in panel B of Table 4 that highly experienced workers (pre-
sumably well informed about their own ability) do not react less at
the revelation stage that highly inexperienced workers.
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1732 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
Nevertheless, it may be that a different model of
learning about absolute ability can explain the reve-
lation stage effect.
3.7.4. Alternative Explanation: Learning About
the Distribution of Performance. One potential ex-
planation of our findings is that the RPIE may have
helped workers to learn about a productivity param-
eter that is common to all workers, and, again, this
may have led to a readjustment in the chosen level of
effort. For instance, a worker with bad luck in a par-
ticular month may, in the absence of RPIE, wrongly
believe that the returns to effort in the company are
not very high. However, by observing the higher pro-
duction of his coworkers, he will update his belief
about the returns to effort upward, which may lead
to more effort.
We believe that this explanation is unlikely. Any
learning explanation is inconsistent with the docu-
mented positive reaction at the kickoff stage, in which
there can be no learning. Second, we would expect
that some workers should react to the arrival of new
information by increasing their effort, whereas others
should decrease it. We find instead in §3.5 that the sign
of the reaction to the RPIE is positive for almost all
Again, it may be that a different model of learning
can explain the revelation stage effect.
3.7.5. Alternative Explanation: Certification to
Outside Employers. It could be argued that the rela-
tive performance information included in the pay slip
could have been used by workers as a verifiable signal
of high productivity to prospective outside employ-
ers. Under this hypothesis, the interest in improving
this signal could have generated the increase in effort
and productivity that we have documented.
Although we cannot completely rule out this
notion, we find it unlikely for two reasons. First, we
would not expect workers at the bottom third of the
distribution to plan on disclosing such a signal to out-
side employers, because little can be gained by admit-
ting one’s relative incompetence.
As a result, we
would expect their effort to remain unchanged, con-
trary to our findings. Second, the availability of such
certification should have led to an increase in the quit
rate among the workforce (and especially among the
best performers), as information about workers’ abil-
ity becomes less asymmetric. We show in §4 that this
is not the case.
This is of course provided that (a) the most that such workers
can aspire to is relative mediocrity and that (b) outside employers
are unaware that our firm provides this information, two notions
that we find highly plausible.
4. Other Effects of the Relative
Performance Information Exercise:
Mistake Rate and Separation Rate
In this section we briefly investigate whether the pro-
ductivity increase documented in the previous section
was accompanied by a permanent change in (a) the
quality of the work done, and (b) the separation rate.
4.1. Effect on the Mistake Rate
The “quality” component represents a tiny determi-
nant of total compensation (on average, the discount
in compensation due to mistakes is less than 1%). Yet,
it is worth studying whether the RPIE was accom-
panied by a permanent increase (or occured at the
expense of a permanent decrease) in the quality of
work done. Unfortunately, in our firm information
on mistakes is recorded only at monthly intervals,
rather than at daily frequencies. Because this makes it
more difficult to study the instantaneous behavioral
response to the introduction of the RPIE, our find-
ings in this Section should be regarded as suggestive,
rather than conclusive.
We display in Figure 4 the evolution of the average
mistake rate over time. The introduction of the RPIE
did not coincide with a dramatic discontinuity in the
evolution of the mistake rate. The mistake rate is, on
average, higher during the 12 months previous to the
RPIE than during the months after it, but this declin-
ing trend is apparent well before the kickoff date. The
month of September 2001 (the first month after the
kickoff date) is associated with a mistake rate that is
much higher than that of the two previous months
(July and August). However, in October, November,
and December the mistake rate was back to normal
pretreatment levels.
Figure 4 Evolution of Mistake Rate Over Time
Kickoff Revelation
Aug99 Sep01 Nov01 Aug02


Notes. The dependent variable is (log of) number of mistakes per good dis-
patched. Vertical dashed lines display September 1, 2001 (kickoff date), and
November 1, 2001 (revelation date). Each dot is the (averaged across work-
ers) mistake rate of a different month.
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
Management Science 57(10), pp. 1721–1736, ©2011 INFORMS 1733
We now use our baseline six month time window
to estimate the following panel data specification:

· o +j
, (4)
where A
is the (log of) mistake rate of worker i on
month |, d
is a dummy variable capturing month |, j
is an individual time-invariant unobserved effect, X

is a vector of covariates similar to that of column (3)
of Table 2, and u
are iid person-specific idiosyncratic
The main parameters of interest are d
, d
, and d
, where the first two
parameters capture the kickoff period, and the second
two parameters capture the revelation period. The
parameter d
is omitted, and the parameter d
captures the pretreatment time trend in the mistake
Table 5 reports ordinary least squares estimates of
(4). In column (1) we do not include worker fixed
effects or extra covariates. We find that in August
the mistake rate is smaller than in July, whereas in
September it is larger. From October onward we do
not find any statistical difference in the mistake rate
relative to the omitted month of July. Column (2)
shows that this finding is robust to conditioning on
individual fixed effects and a vector of covariates sim-
ilar to that of column (3) in Table 2.
Thus, the results in this section are consistent with a
transitory (i.e., in the month of September) increase in
the mistake rate following the kickoff date. There is,
however, no evidence to suggest that the RPIE led to
a permanent change in the quality of work produced.
4.2. Effect on Workers’ Separation Rate
In this section we investigate whether the separa-
tion rate can be regarded as being different before
and after the introduction of the RPIE. We also study
whether workers of different ability reacted differ-
ently to this introduction. To do this, we estimate
a Cox semiparametric proportional hazard model,
which allows for a fully flexible, nonparametric base-
line hazard. In our model, the covariates shift the
baseline hazard proportionally, through the function
· ¹
) +X

· o, (5)
where í
=1 if the first date of the quarter of tenure |
occurs after the revelation moment, ¹
is a measure
of individual ability,
and X

is a vector of worker
To compute this measure, we run a regression where productivity
(i.e., the number of goods per hour) is the dependent variable, and
day fixed effects, worker fixed effects, and quarter of experience
dummies are the explanatory variables. Our measure of individual
ability, ¹
, is the estimated worker fixed effect from this regression.
The estimated standard errors are bootstrapped in Equation (5) to
account for the fact that ¹
is generated, rather than observed.
Table 5 Evidence on the Mistake Rate
(1) (2)
Unconditional Controls
July (baseline month 1) Omitted Omitted
August (baseline month 2) −0.167

(0.076) (0.112)
September (kickoff month 1) 0.192
(0.069) (0.081)
October (kickoff month 2) 0.017 0.114
(0.072) (0.083)
November (revelation month 1) 0.036 0.143
(0.063) (0.118)
December (revelation month 2) −0.056 −0.016
(0.080) (0.112)
Worker fixed effects No Yes
Table 2 covariates No Yes
Adjusted l
0.03 0.70
Number of observations 332 332
Notes. Dependent variable is (log of) number of mistakes per good dis-
patched. Data comprise July 1, 2001, to December 28, 2001. July and August
are the two months previous to the introduction of the RPIE. September and
October are the two months comprising the kickoff period. November and
December are the first two months of the revelation period. The sample is
restricted to those workers with at least six months of experience in the firm.
Robust standard errors allow for cluster at the worker level. Covariates are
as in column (3) of Table 2 (except for the linear time trend and the (log of)
goods piece rate). The covariates are monthly averages.

, and
denote estimates significant at 10%, 5%, and 1%,
We estimate this model exclusively on workers
who joined the firm after the end of the probation-
ary period (i.e., for whom the quarter of tenure | is
at least 3). By focusing on workers past their proba-
tionary period we can interpret our findings as refer-
ring to the workers’ quit rate.
The small size of the workforce, together with the
fact that separations are relatively rare events, makes
it unfeasible to circumscribe the analysis to a narrow
time window centered around the introduction of the
RPIE. We instead use eight years of data, between
February 1999 and December 2007. Because of the
large width of the sample time window, the evidence
from this section should again be regarded as sugges-
tive, rather than conclusive.
We include in our analysis only workers who joined
the firm after August 1999 (the first month for which
we have observations) to avoid problems of left cen-
soring. In total, our sample contains 127 workers, out
of which 41 had left the firm by December 2007.
Table 6 reports the results from the proportional
Cox hazard model (5). In column (1) we observe that
the likelihood of a separation is lower in the period
after the introduction of the RPIE than in the period
before. The hazard ratio is 0.181, indicating that the
likelihood of a separation in the period after the RPIE
is just 18% of the likelihood in the period before. In
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1734 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
Table 6 Likelihood of a Separation Before and After the Revelation of
Relative Performance
Dependent variable: equal to 1 if worker i left the firm in quarter t
(1) (2) (3)
Cox regression Cox regression Cox regression
unconditional ability other covariates
Revelation −1.70
(0.359) (0.367) (0.537)
|0.181] |0.178] |0.193]
Worker ability −6.24
(1.19) (1.42)
|0.001] |0.0003]
Revelation × −0.008 0.216
Worker ability (0.491) (0.489)
|0.486] |1.24]
Worker covariates No No Yes
Number of observations 719 719 719
Number of subjects 127 127 127
Number of separations 41 41 41
Notes. Revelation takes value 1 if the first date of quarter t occurs after
November 1, 2001. Worker ability is the estimated worker fixed effect of a
regression that has productivity as a dependent variable and includes expe-
rience dummies and day fixed effects, as well as worker fixed effects, on the
right-hand side. Worker covariates include gender, marital status, age entry
in the firm, and a linear cohort effect (the calendar year in which the worker
joined the firm). Data comprise August 1, 1999, to December 31, 2007. The
sample is restricted to workers joining the firm after August 1999. The sam-
ple is restricted to workers surviving the probationary period (i.e., the first six
months of employment). Robust standard errors are in parentheses in col-
umn (1). Robust bootstrapped standard errors are in parentheses in columns
(2) and (3). Hazard ratios are in square brackets.

, and
denote estimates significant at 10%, 5%, and 1%,
column (2) we introduce our measure of worker’s
ability both in levels and interacted with í
. We
find that workers with a higher value of our ability
measure are less likely to leave the firm, suggesting
that our measure does indeed capture an important
dimension of worker heterogeneity. However, the dif-
ference in separation rates across workers with differ-
ent abilities does not change after the introduction of
the RPIE. Last, we find in column (3) that our find-
ings are robust to the introduction of a number of
worker characteristics, including gender, marital sta-
tus, age at entry in the firm and a linear cohort effect
(captured by the calendar year in which the worker
joined the firm).
To summarize, we do not find any evidence indicat-
ing that workers are more likely to quit the firm after
the introduction of the RPIE. Subject to the caveats
mentioned above, our evidence in fact suggests the
4.3. Relation with Monetary Incentives and
Effect on Profits
We now relate the effects of the RPIE to the effects
of monetary incentives. To estimate how workers
respond to an increase in the goods piece rate, we use
data between January 2002 and December 2007.
regress the (log of) individual productivity on the (log
of) the goods piece rate, the (log of) the orders piece
rate (which the firm changed during this time period),
a cubic time trend, individual fixed effects, and the
time-varying controls from Table 2. Our estimate of
the effect of increasing the goods piece rate implies an
elasticity of 43%. This elasticity implies that the 6.8%
achieved by the introduction of the RPIE would be
equivalent to an increase in the goods piece rate of
around 15.8%, equal to 6.8%,0.43. Thus, the nonpecu-
niary incentives provided by the RPIE are reasonably
large not only in absolute terms but also relative to
the effects achieved by monetary incentives.
We can also calculate the potential reduction in
labor costs induced by the RPIE. Because the existing
65 workers started producing on average 6.8% more
goods per hour, the firm was in principle able to pro-
duce the same amount of goods in the same amount
of time with 4.20 (6.5%) fewer workers. How would
this have translated into labor cost savings? To calcu-
late this, we need to take into account the fact that
the performance component of worker compensation
is independent of the number of workers in the firm.
The performance component is given by the prod-
uct of the piece rate(s) and total production, which
is exogenously given by the demand for the firm
products. The firm could, however, have saved on
the fixed component of worker compensation, which
represented around 75% of the average worker pay
during this period. Producing the same amount with
6.5% fewer workers therefore implies potentially sav-
ing 4.9% = 6.8% × 75% in terms of labor costs. In
our opinion, this represents a economically signifi-
cant, although not huge, amount.
5. Concluding Remarks
We have provided evidence that social comparisons
affect human behavior independently of the mone-
tary incentives to which these comparisons are usu-
ally linked.
We use this longer time period because the fact that the firm
changed the level of the piece rate on repeated occasions and by
significant amounts allows us to obtain good estimates of the effects
of monetary incentives.
We refer to these savings as potential savings rather than actual
savings, because no worker was actually fired as a result of the
increase in productivity. The firm does not typically fire employ-
ees past the probationary period, and during this time period the
demand for the firm products and the size of the workforce was
growing over time. The firm therefore adjusted by failing to expand
the workforce at the same rate as it would have in the absence
of the RPIE. Note, finally, that the decrease in the separation rate
found in §4.2 will also reduce search costs and therefore increase
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
Management Science 57(10), pp. 1721–1736, ©2011 INFORMS 1735
In our firm, simply receiving information about rel-
ative standing—and anticipating such future informa-
tion—leads to an increase in individual productivity
that is statistically and economically significant. We
have found no evidence of this increase being purely
transitory. Apart from a transitory (i.e., one month)
slight decrease in the quality of work done, we have
found no evidence of workers reacting along other
dimensions in ways that are detrimental to the firm.
Taken together, our results suggest that making infor-
mation about relative performance available is unam-
biguously beneficial to the firm’s profitability.
We believe that the notion that workers display rel-
ative concerns provides the best possible explanation
of our findings. Other interpretations fall short in one
way or another, notably in explaining why workers
already increase their effort at the kickoff stage.
We have been able to identify the causal effects
of social comparisons thanks to unique access to the
detailed personnel records of a firm that, for largely
exogenous reasons, started to provide information on
relative performance to workers. Our findings are
based on a single firm study, and may not necessar-
ily generalize to other settings. To understand their
external validity, there are two features of our firm
and work environment that we regard as particularly
(1) The effects of providing information on rela-
tive productivity and pay may depend on the tech-
nology of production, the incentive scheme in place,
and the nature of workers affected. In our firm, the
existence of absolute performance pay implies a one-
to-one mapping between performance and compen-
Furthermore, work is individual, with little
scope for cooperation among workers. As a result,
workers dissatisfied with their relative position have
one single natural channel through which they can
remedy their situation: the exertion of more effort.
The effects of information on social comparisons may
obviously be different in fixed-wage settings or in set-
tings where industrial politics are potentially impor-
tant (Lazear 1989). Similarly, workers’ attitudes may
be critical in determining whether social comparisons
are morale boosting or have the opposite effect. In
short, although we may find similar effects among,
say, door-to-door salesmen, our results do not neces-
sarily generalize to, for instance, corporate lawyers,
because for these team-work is important and there is
large scope for mutual sabotage.
(2) Perhaps more importantly, in our firm the infor-
mation was disclosed in a purely private manner to the
Obviously this one-to-one mapping also implies that our results
may reflect workers caring about relative productivity per se, as
well as about relative pay. Disentangling these two possibilities falls
outside the scope of this paper.
workers (although, of course, some may have shared
that information among themselves). We are unfor-
tunately unable to judge the extent of information
sharing among workers. However, to the extent that
such information sharing was not complete, we may
observe different, possibly stronger, results under a
public disclosure of relative standing. Falk and Ichino
(2006), for instance, find that the ability to observe
other workers’ productivity leads to an increase in
effort in a flat wage setting. Although the distinction
between the impact of private and public information
is clearly beyond the scope of this paper, we regard it
as an interesting direction for future work.
The authors thank David de Meza, Daniel Ferreira, Thomas
Kittsteiner, Ignacio Palacios-Huerta, and Yona Rubinstein
for insightful comments. They are grateful to all those
involved in providing the data. Joy Blundell and Vera
Radyugina provided excellent research assistance. This
paper has been screened to ensure that no confidential
information is revealed. All errors remain the authors’
Anderson, R., S. Manoogian, S. Reznick. 1976. The undermining
and enhancing of intrinsic motivation in preschool children.
J. Personality Soc. Psych. 34(5) 915–992.
Azmat, G., N. Iriberri. 2010. The importance of relative perfor-
mance feedback information: Evidence from a natural exper-
iment using high school students. J. Public Econom. 94(7–8)
Bandiera, O., I. Barankay, I. Rasul. 2009. Team incentives: Evidence
from a field experiment. Working paper, London School of Eco-
nomics, London.
Bandura, A., D. Cervone. 1983. Self-evaluative and self-efficacy
mechanisms governing the motivational effects of goal sys-
tems. J. Personality Soc. Psych. 45(5) 1017–1028.
Baron, R. 1988. Negative effects of destructive criticism: Impact on
conflict, self-efficacy and task performance. J. Appl. Psych. 73(2)
Buunk, A., F. Gibbons. 2007. Social comparison: The end of a theory
and the emergence of a field. Organ. Behav. Human Decision
Processes 102(1) 3–21.
Casas-Arce, P., F. A. Martínez-Jerez. 2009. Relative performance
compensation, contests, and dynamic incentives. Management
Sci. 55(8) 1306–1320.
Charness, G., P. Kuhn. 2007. Does pay inequality affect worker
effort? Experimental evidence. J. Labor Econom. 25(4) 693–725.
Clark, A., P. Frijters, M. Shields. 2008. Relative income, happiness
and utility: An explanation for the easterlin paradox and other
puzzles. J. Econom. Literature 46(1) 95–145.
Clark, A., D. Masclet, M. Villeval. 2010. Effort and comparison
income: Experimental and survey evidence. Indust. Labor Rela-
tions Rev. 63(3) 407–426.
Deci, E. 1972. Effects of contingent and noncontingent rewards and
controls on intrinsic motivation. Organ. Behav. Human Perfor-
mance 8(2) 217–229.
Deci, E. 1975. Intrinsic Motivation. Plenum Press, New York.
Deci, E., R. Koestner, R. Ryan. 1999. A meta-analytic review of
experiments examining the effects of extrinsic rewards on
intrinsic motivation. Psych. Bull. 125(6) 627–668.
Blanes i Vidal and Nossol: Tournaments Without Prizes: Evidence from Personnel Records
1736 Management Science 57(10), pp. 1721–1736, ©2011 INFORMS
Delfgaauw, J., R. Dur, J. Sol, W. Verbeke. 2009. Tournament incen-
tives in the field: Gender differences in the workplace. Working
paper, Erasmus University, Rotterdam, The Netherlands.
Eriksson, T. 1999. Executive compensation and tournament theory:
Empirical test on danish data. J. Labor Econom. 17(2) 262–280.
Eriksson, T., A. Poulsen, M. C. Villeval. 2009. Feedback and incen-
tives: Experimental evidence. Labour Econom. 16(6) 679–688.
Ertac, S. 2006. Social comparisons and optimal revelation: The-
ory and experiments. Working paper, Koç University, Istanbul,
Etzioni, A. 1971. Modern Organizations. Prentice-Hall, Englewood
Cliffs, NJ.
Falk, A., A. Ichino. 2006. Clean evidence on peer effects. J. Labor
Econom. 24(1) 39–57.
Fehr, E., S. Gächter. 2000. Fairness and retaliation: The economics
of reciprocity. J. Econom. Perspect. 14(2) 159–181.
Fehr, E., K. Schmidt. 1999. A theory of fairness, competition and
cooperation. Quart. J. Econom. 114(3) 769–817.
Fershtman, C., H. Hvide, Y. Weiss. 2005. Cultural diversity, status
concerns and the organization of work. Res. Labour Econom.
24(1) 3–38.
Festinger, L. 1954. A theory of social comparison processes. Human
Relations 1(7) 117–140.
Frank, R. 1984. Are workers paid their marginal products? Amer.
Econom. Rev. 74(4) 549–571.
Frank, R. 1985. Choosing the Right Pond: Human Behavior and the
Quest for Status. Oxford University Press, New York.
Kluger, A., A. DeNisi. 1996. The effects of information interventions
on performance: A historical review, a meta-analysis and a
preliminary information intervention theory. Psych. Bull. 119(2)
Knoeber, C., W. Thurman. 1994. Testing the theory of tournaments:
An empirical analysis of broiler production. J. Labor Econom.
12(2) 155–179.
Kohn, A. 1993. Punished by Rewards. Houghton-Mifflin, Boston.
Kuhnen, C., A. Tymula. 2010. Feedback, self-esteem and perfor-
mance in organizations. Working paper, Northwestern Univer-
sity, Evanston, IL.
Latham, G., K. Wexley. 1981. Increasing Productivity Through Perfor-
mance Appraisal. Addison-Wesley, Reading, MA.
Lavy, V. 2008. Gender differences in market competitiveness in a
real workplace: Evidence from performance-based pay tourna-
ments among teachers. NBER Working Paper 14338, National
Bureau of Economic Research, Cambridge, MA.
Lazear, E. 1989. Pay equality and industrial politics. J. Political
Econom. 97(3) 561–580.
Lazear, E., S. Rosen. 1981. Rank-order tournaments as optimum
labor contracts. J. Political Econom. 89(5) 841–864.
London, M. 2003. Job Feedback: Giving, Seeking and Using Feedback
for Performance Improvement. Lawrence Erlbaum Associates,
Mahway, NJ.
Luttmer, E. 2005. Neighbors as negatives: Relative earnings and
well-being. Quart. J. Econom. 120(3) 963–1002.
Moldovanu, B., A. Sela, X. Shi. 2007. Contests for Status. J. Political
Econom. 115(2) 338–363.
Moore, D., W. Klein. 2008. Use of absolute and comparative perfor-
mance feedback in absolute and comparative judgments and
decisions. Organ. Behav. Human Decision Processes 107(1) 60–74.
Nadler, D. 1979. The effects of feedback on task group behavior:
A review of the experimental research. Organ. Behav. Human
Performance 23(3) 309–338.
Oswald, A. 1997. Happiness and economic performance. Econom. J.
107(6) 1815–1831.
Sunde U., T. Dohmen, A. Falk, B. Weber, K. Fliessbach. 2011. Rela-
tive versus absolute income, joy of winning, and gender: Brain
imaging evidence. J. Public Econom. 95(3) 279–285.
Tran, A., R. Zeckhauser. 2009. Rank as an incentive. Working paper,
Indiana University, Bloomington.